Artificial intelligence is changing the way data centers are built: these are the challenges and trends in data centers

Artificial intelligence (AI) is not new, but the expansion and democratization of its use is changing the way both companies and employees work, as well as how people interact. Its impact goes even further. In fact, it is having significant consequences for data centers, not only in terms of their operational capacity to meet the demands of running various AI-based applications and solutions, and the wave of innovation associated with this technology, but also in the construction and design of data centers.

In fact, with the increased use of AI, data centers must adapt and prepare to support a greater demand for GPUs, as well as power and cooling . And all this without neglecting sustainability and operational efficiency .

"Artificial intelligence implies a change in design and construction due to the higher density per rack. In addition, there's the thermal challenge of cooling the data center more effectively," Tomás Callejo, director of /fdata, Fibratel's data center unit, told elEconomista.

These changes are primarily due to the widespread adoption of artificial intelligence models and machine learning applications, which are driving demand for GPUs , high-power servers, and higher-density racks . "AI impacts operations, the cloud , and the construction of larger and smaller machines," he adds.

These factors necessitate a redesign of the architecture of facilities where the challenge lies in meeting these energy needs. According to estimates from the Uptime Institute , industry analysts calculate that energy consumption will increase between 10% and 15% year-on-year, making it essential to move towards a more efficient electricity distribution network with greater generation capacity . This is one reason why there is a growing commitment to renewable energy. In Spain, unlike other countries such as the United States, there are currently no applications to use nuclear energy through the installation of mini-power plants.

But it also requires addressing the increased cooling needs in the form of advanced air conditioning solutions, which account for 20%, and liquid cooling (which uses a liquid—usually water—to dissipate heat from electronic components), representing 80%. In the latter case, these are closed-loop water systems, which are the preferred option in Spain, since "they only consume water upon arrival," adds Callejo.

Another change that will become established is related to modular data centers due to their scalability, flexibility, and rapid deployment at low cost, as well as micro data centers , which are emerging as a good option for use cases that require low latency and local autonomy, such as Internet of Things (IoT) projects, Edge computing (data processing that takes place on peripheral devices or servers located as close as possible to the data source), or remote offices, among other examples.

Another challenge facing data centers, impacting their design and construction, is sustainability . This is precisely one of the reasons why data centers are under scrutiny. However, "data centers were designed to be sustainable from the start. They will all strive to use renewable energy, as well as utilize waste heat for other purposes ," which even extend beyond the data center itself, such as heating municipal swimming pools or the homes in the surrounding area. These uses are already being implemented in Europe and are expected to be adopted in Spain in the future.

But it's also another trend that will be observed in data centers because carbon footprint is becoming a strategic KPI . This also explains why data centers are already opting for renewable energy sources, heat recovery systems, optimized water usage, and monitoring tools to measure consumption in real time. " Sustainability policies must also extend to the technological infrastructure ," so that workload utilization and overall performance are maximized.

Another major change in data centers, and a challenge in itself, is their evolution toward more autonomous and resilient environments , resulting in reduced operating costs and improved availability. This trend will be driven by the integration of AI and advanced analytics to reduce failures, optimize maintenance, perform predictive maintenance, and adjust capacity based on demand.

Another trend worth mentioning is talent . In addition to ICT and artificial intelligence profiles applied to operations, technical profiles are also needed in areas such as advanced refrigeration solutions and energy management . In this context, internal training and collaboration with universities and technology centers are emerging as key to having qualified professionals and ensuring generational renewal.

These changes and challenges are driving an " unprecedented transformation in the sector," notes Callejo, one that goes beyond simply redefining the architecture and management of data centers. In fact, only those organizations capable of anticipating these changes with more efficient, sustainable, and automated infrastructures "will be able to maintain their competitiveness in the coming years."